Ecology,68(3), 1987, pp. 478-490 ? 1987by the Ecological Society of America DYNAMICS OF MOJAVE DESERT SHRUB ASSEMBLAGES IN THE PANAMINT MOUNTAINS, CALIFORNIA1 Robert H. Webb, John W. Steiger, and Raymond M. Turner United States Geological Survey, 300 West Congress Street FB-44, Tucson, Arizona 85701 USA Abstract. We studied shrub communities in the Panamint Mountains of the Mojave Desert to determine whether vegetational changes after disturbance can be characterized as "succession" according to modern successional theory. We found, on a variety of dis- turbed and undisturbed sites, that the rate of change was a function of the type and age of disturbance. Recent debris-flow deposits were colonized by shrub assemblages of different species composition than those on the surrounding, older debris-flow deposits and other geo- morphically stable surfaces. Colonization of human-disturbed sites was highly variable, but species compositions were different from the predisturbance species composition. In Grayia-Lycium assemblages, Grayia spinosa reasserted its dominance over colonizers rel? atively quickly. In Coleogyne assemblages, typically found on older geomorphic surfaces, species composition differences persisted considerably longer, depending on the severity of the initial disturbance. Extremely stable assemblages, dominated by Coleogyne ramo- sissima, occurred on the oldest, least disturbed surfaces. The variability of species com? position among disturbed sites was greater than the variability among undisturbed and geomorphically stable sites, in accord with "convergent succession." Models of desert succession should consider several factors: (1) colonization is depen- dent largely on the severity of disturbances and residual biotic components; (2) the time span for recovery may be longer than past periods of climatic and geomorphic stability; and (3) colonizing species may have considerable range in their life-history strategies. Key words: Coleogyne ramosissima; convergent succession; desert shrubs; disturbance; ghost towns; Grayia spinosa; Lycium andersonii; Mojave Desert. Introduction Despite these fluctuations (18-69% turnover), the rel? ative contribution of species remained similar. Species Desert plant communities (sensu Daubenmire 1968) were functionally to traits have long been recognized as sensitive to disturbance. grouped according life-history of survivorship, size, recruitment, and The cliche "easily scarred, slowly healed" implies that longevity, age structure. The endpoints in this continuum of the rate of revegetation of disturbed desert areas is slow strategy were with life compared with human life spans. The questions ofthe species long spans, high survivorship, size, and structures skewed toward older in? rate and course of revegetation have been addressed large age dividuals; and with short life low sur? in several studies of plots disturbed up to 66 yr prior species spans, small size, and structures skewed toward to measurement (Wells 1961, Vasek et al. 1975<2, b, vivorship, age individuals. Their a hierarchical Webb and Wilshire 1979, Lathrop and Archbold 1980a young study suggests structure that affect in b, Vasek 1980, Prose and Metzger 1985, Carpenter et life-history may revegetation Sonoran Desert communities. al. 1986). These studies found that revegetation re? Upland In another measured fluctua? quires longer than 66 yr, is dependent on the severity study, Beatley (1980) tions in cover of shrubs in the northern of disturbance, and does not necessarily follow a pre- perennial Desert. turnovers caused dictable model of species-composition change. Mojave Large by recurring resulted in 14% death and a 20-30% The stability of desert shrub communities has re- wet-dry cycles recruitment either or propa- ceived little attention, possibly because of an erroneous (by seedlings vegetative rate over a period. However, neither perception of lack of change (West 1982). Goldberg gation) 12-yr composition nor the relative proportion of each and Turner (1986) studied permanent plots in the So- species concluded that fluctuations in noran Desert measured at irregular intervals over 72 species changed. Beatley the cover of desert shrub communities are closely cor? yr and found large fluctuations in cover and density. related with climate but are low enough to allow con? tinued stability. 1 Manuscript received 4 June 1986; revised 23 August 1986; The results of revegetation and stability studies have accepted 27 August 1986. created a polarization of opinion on whether plant June 1987 MOJAVE DESERT SHRUB COMMUNITY DYNAMICS 479 succession occurs in deserts. Shreve (1942) denied any species composition change following disturbances and observed virtually no biotic control of substrates or revegetation processes. Recent studies (Vasek and Lund 1980, Vasek 1980, 1983, West 1982) have challenged Shreve's assertions based on perceived seral stages as? sociated with soil chemical changes resulting from pedogenesis. Shreve's proponents (e.g., Lathrop and Rowlands 1983) cite the absence of soil formation in deserts as a weakness in desert succession models. Moreover, while colonizers of disturbed areas may have a different species composition from the undisturbed vegetation (Wells 1961), they usually have some low- level representation in the undisturbed assemblages (Beatley 1976), blurring the distinction among seral stages. MacMahon (1981) refers to the "conspicuous- ness" of succession as increasing with the predictability of rainfall, and his model, which spans North Amer? ican biomes, affords a means of reconciling views of succession in humid vs. arid areas. But a concensus as to the applicability of succession theory to deserts has not been achieved. Part of the problem stems from iack of field data to evaluate either rates of recovery from disturbance or long-term dynamics of shrub as? semblages. The Panamint Mountains of California offer an ideal setting for the evaluation of dynamics of Mojave Des? ert shrub communities. Debris flows periodically dis- turb vegetation, forming a mosaic of geomorphic sur? Fig. 1. Location map for the Panamint Mountains study faces of different ages. Humans have also disturbed sites, California. vegetation periodically over the last 79 yr, leaving sites with varying degrees of revegetation. In some cases, human disturbance has crossed different-aged geo? 26?C, respectively, from a 10-yr record. From regional morphic surfaces. The purpose of this paper is to in? precipitation records, Hunt et al. (1966) estimated an vestigate the successional dynamics for two shrub com? annual precipitation range from 50 mm at the base to munities, and to relate the results to modern 380 mm at the top ofthe Panamint Mountains. The successional theory. study areas reported here probably have an annual precipitation of 200-250 mm, most of which falls in Methods winter months. Setting Vegetation assemblages The Panamint Mountains are a north-south trending We studied the two upper elevation Mojave Desert range bordering the west side of Death Valley (Fig. 1 community types of Grayia-Lycium and Coleogyne inset). Elevations range from ?100 m on the Death (Beatley 1976) lying between 1530 and 1830 m ele? Valley side and +320 m on the Panamint Valley side, vation. The presence/absence of Coleogyne ramosis- to 3550 m on Telescope Peak. The sites used in this sima, with a variable lower elevation limit of 1580? study lie in an upland area of low hills and alluvial 1700 m, is the distinguishing floristic feature between valleys at elevations of 1530-1830 m. The bedrock these two communities, because both Grayia spinosa geology consists of Cambrian and Precambrian cal- and Lycium andersonii occur in Coleogyne assem? careous sedimentary rocks intruded with Tertiary gran- blages. Grayia-Lycium communities have not been re? ite along the axis of the range (Hunt and Mabey 1966). ported in either the Black or the Grapevine Mountains The principal rocks in the study areas are the Tertiary on the east side of Death Valley (Kurzius 1981, granitic intrusions and Precambrian Noonday Dolo- Schramm 1982), but occur in the Cottonwood Moun? mite and underlying Kingston Peak Formations. tains, the northerly extension of the Panamint Range Climate of this area is poorly documented. The only (Peterson 1984). Coleogyne does not occur in the climatological station in the range, at Wildrose Canyon Grapevine Mountains (Kurzius 1981) but forms ex- (1250 m), has a mean annual precipitation of 191 mm tensive stands in the Black Mountains (Schramm 1982) and mean January and July temperatures of 4? and and in the Cottonwood Mountains (Peterson 1984). 480 ROBERT H. WEBB ET AL. Ecology, Vol. 68, No. 3 Vegetation measurements We used repeat photographs to help determine sites for and to assess Vegetation measurements were made on debris-flow vegetation measurement, qualitatively the relative and population turnover of surfaces in Wood Canyon, in the ghost towns of Skidoo longevity and Harrisburg and adjacent undisturbed sites, and species. along selected segments ofthe abandoned Skidoo pipe- line (Fig. 1). Choice of these sites provided a compar? Results: Dynamics of Coleogyne Communities ison of colonization on naturally disturbed surfaces Debris flows in Wood Canyon with revegetation of townsites, and both can be com? pared with the natural fluctuations in undisturbed Wood Canyon (Fig. 1) has a surficial geology com- Grayia-Lycium and Coleogyne communities. Study posed of modern and ancient debris flows on alluvial sites were chosen